Study claims Live TV piracy growing (wrongly)

A study by Google and PRS for Music is suggesting that live TV streams is currently the fastest-growing aspect of online piracy.  The study identified 153 sites "believed to be significantly infringing copyright" to see the kinds of content and services offered, and the business models used.  Researchers identified six basic types of core activities on the sites - Live TV Gateway, Peer-to-Peer (P2P) Community, Subscription Community, Music Transaction, Rewarded Freemium, and Embedded Streaming.  They then examined a further 104 sites to validate the appropriateness of the six core activities.

  The study found that Live TV Gateways (sites offering links to streams of live free-to-air or pay-TV channels) accounted for roughly a third of the sample, and was the fastest growing areas of core activity.  Advertising support was the primary business model for two thirds of those sites in the sample, although the report claimed that the majority of ads were for companies "outside of the mainstream."  The report suggested that many of these sites also solicited donations from users as part of their business model.
  It should be noted that if these sites only offered links to legally licensed TV streams (say to network, station, or program sites) and not links to unlicensed sites or unlicensed content, that activity does not fit current definitions of online piracy.  As such, statements or inferences that such activity is piracy, or that all TV streaming gateways are illicit or an area of online piracy is not necessarily accurate.
  Peer-to-Peer Communities were identified as the second fastest growing segment, and relied even more heavily on advertising for revenues.  Of the websites in the sample with P2P sharing as their core activity, 86% carried advertising.  In contrast, subscription and download sites used user payment systems as a revenue source.  Some 69% of such sites featured credit card logos, and 39% had separate payment pages. The sites in the sample offered a wide range of of digital content, from films and music to games and ebooks.

Google's Theo Bertram said that "The evidence suggests that one of the most effective ways to do this is to follow the money, targeting the advertisers who choose to make money from these sites and working with payment providers to ensure they know where their services are being used."

PRS for Music chief executive Robert Ashcroft added: "This groundbreaking research tells us two things. Firstly sites involved in copyright infringement are businesses with real costs and revenue sources. They receive subscription or advertising revenue, pay their server or hosting costs but fail to pay the creators of content on which their businesses depend.
"Secondly, not all of these business models are the same, and the government now has the evidence to understand which policy levers to apply to deal with these different businesses effectively."

  You have to be careful with a lot of these "rampant piracy" studies, as they tend to overestimate both the frequency and impact of the alleged piracy.  So I took a look at the actual study and methodology, and will toss in my own warnings about the validity of some of the claims.
  First, I could find no indication that the researchers actually checked to see if the content and services offered by the sites were illicit. They only measured if content of a certain type of content was present on the site or not, and for some content types how many of the top ten examples of that content were available.  There was no measure of whether the content was offered legally, or whether content owners were paid for access to their content by the sites.
  Second, the study did not directly observe business models of site owners, but rather looked for evidence of features assumed to be associated with certain types of business models.  That is, they did not directly measure a site's commercial motives or choice of business models (or the success of any such models).  Rather, the choice and relative importance of a business model to a site is inferred.  The study thus can not validly claim that these are the models, or which are predominant - only that sites seem to be employing some aspects of particular business models.
  Third, all revenues and costs were derived indirectly from specific formulas, and in many cases, from assumed values.  For example, revenues were estimated by multiplying the number of page views times the lowest listed price for that type of content.  The formula assumes all page visits result in sales, and that there is no "free" content available - both highly unlikely assumptions.
  Fourth, this was a self-selected sample of sites identified by UK "experts" and is not generalizable, either as sites "involved in copyright infringement," or more broadly as digital content access sites.  The study methodology explicitly assumes the sites engage in online piracy, but offers no evidence in support for that assumption.  Any "finding" involving piracy is thus not a valid finding of the study, but merely an assumption.  And the purposive sampling method precludes the randomness needed for generalization to wider populations.  Any study findings are valid only as descriptions of that particular sample.

  These issues don't necessarily invalidate the first parts of the study - identifying the types of site content and services offered on the sampled sites, and the range of business models they seem to be employing.  The study can and does validly claim that there seem to be differences in the types of business models  the sampled sites seem to be using, and that the segmenting of content and services for the sampled sites are a reasonable way of distinguishing types of content/services (if not the only possible way).  However, any specific measures and proportions aren't generalizable to the larger population of websites (either legal or pirate), and the actual commercial motives business models are inferred rather than observed. In addition, any revenues or cost estimates derived must be considered to be unreliable and imprecise (at best).  But most important from a validity perspective is the fact that the sites are only suspected of being online content pirates (or even of being commercial).  The study did not directly confirm either presumption.  Specifically, the first key "finding" identified by Roger Ashcroft above is not valid findings of the study, and would have been roundly rejected as valid conclusions by any serious peer review.
  In particular, the definition and description of the "TV Gateway" type sites make it clear that these are quite different from other content-sharing or downloading sites - and in fact, sites that only link to other places where content is legally available aren't necessarily violating copyright or engaged in piracy.  Any attempt to broadly label TV gateway services as pirates is misleading at best, and wrong and libelous at worst.  The number of such sites, and even their use, may be growing, but the study fails to demonstrate that such sites are engaged in online piracy.  Journalists should be very cautious of repeating such claims

Sources - Live-TV fastest-growing area of online piracy, says study,  Digital Spy
The six business models for copyright infringement, PRS for Music/Google report